This invention relates to the field of geophysical exploration using measurements of naturally occurring or artificially induced geomagnetic fields and induced geoelectric (telluric) fields, collectively called magnetotelluric fields.
The flow of the telluric currents through the earth's crust depends upon the conductivity or resistivity of the structure of the crust at any particular point. If this conductivity or resistivity can be measured and mapped, information about the structure, and in particular as it may relate to the presence of hydrocarbon, mineral or geothermal resources, can be gained. This is particularly useful in areas where seismic methods of geophysical surveying cannot be used, such as where the sedimentary rocks are overlaid with a thick volcanic layer.
The telluric currents constantly vary in magnitude, direction and polarity, and constitute a complex spectrum of components with various frequencies.
The first practicable electromagnetic survey method was described by Cagniard in U.S. Pat. No. 2,677,801. This consists of measuring and recording over a period of time the variations in one horizontal component of the telluric field, and simultaneously measuring the recording the orthogonal component of the geomagnetic field. These measurements are then converted into frequency components by means of Fourier analysis. The ratio of the frequency component of the electrical field to that of the magnetic field is a wave impedance that is a function of frequency. Since the depth of penetration of an electromagnetic wave into the earth is inversely related to the frequency of the wave and the conductivity of the earth, the wave impedance can be used to estimate the conductivity distribution in a vertical direction through the earth's surface. Cagniard made this estimate using a mathematical model in which conductivity varied only with depth, the so-called 1-D model.
This method was subsequently developed by other workers to be used with models in which the conductivity varies with one horizontal coordinate as well as with depth, the 2-D model. The coordinate along which conductivity is constant in this model is called "strike". In this model there are two cases to be considered, based on the polarization of the magnetotelluric fields. In these two cases the electric fields are polarized respectively parallel and perpendicular to strike. Theoretical studies show that the E parallel impedance function can be used with the techniques developed for 1-D to give reasonably accurate estimates of the conductivity directly beneath the site at which the function is computed. However, the E perpendicular impedance functions have very different properties, and cannot be directly inverted to produce an estimate of conductivity which is of reasonable accuracy. The most common solution of this problem is to use the curves obtained at more than one site, and to adjust the conductivity distribution in the model by an iterative process until theoretical E perpendicular curves which best fit the observed results are obtained. This method has many drawbacks, in particular it requires a considerable amount of computational effort, and there is no assurance that the solution obtained is in fact the correct one.
These methods work well when the area being surveyed does have conductivity which varies only in one or two directions. Unfortunately, however, such two-dimensional variation is rare. When magnetotelluric measurements are made over a structure having a three dimensional conductivity distribution, the following problems are encountered. First, it is generally not possible to identify a set of principal axes as it is in the 2-D case, although a number of ad hoc procedures exist to generate a set of axes. Second, whatever coordinate system is used, it is not possible to separate the electric field into the two distinct cases with the distinctive properties of E parallel and E perpendicular found in the 2-D case. The desirable properties of the E parallel case disappear first, and both cases take on properties similar to those of E perpendicular. Various ad hoc procedures were developed to try and deal with this problem, but they have had limited success. A third problem is that the considerable additional complexity significantly increases the difficulty of ensuring that the interpretation of the data obtained is in fact the correct one.
One way of reducing the third problem is to take many more measurements. In the classical magnetotelluric method, it is usually more difficult to measure the magnetic field than it is to measure the electrical field, making a large number of measurements time-consuming and expensive. However, it is an observed fact that the horizontal components of the magnetic field typically vary much more slowly over distance than do the horizontal components of the electric field. This gave rise to the magnetotelluric-telluric method, which involves the measurement of two orthogonal components of the magnetic field at a limited number of sites distributed over the survey area, and measurement of two orthogonal components of the electrical field at many more sites. The data are then processed using techniques similar to those described for the magnetotelluric model. However, when the conductivity distribution is three dimensional, the results are still found to be fairly unreliable. One variant of this method, which uses a method equivalent to the "roll-along" method of seismic geophysical exploration, is described in U.S. Pat. No. 4,286,218. Overlapping measurements of the. electrical fields along the survey line improve the signal to noise ratio, and the magnetic field is measured at either end of the line, the intermediate magnetic field at the points at which the electrical field is measured being derived by interpolation.
The main disadvantage of all these conventional methods is that they give very unreliable results when the electrical and magnetic characteristics of the subsurface structure vary in all three dimensions. Prior attempts to overcome this unreliability have involved huge computational efforts, which are rarely justified in terms of the results produced.
Another disadvantage of conventional magnetotelluric survey methods is that two measurements of the electrical field, preferably in orthogonal directions, are required at each survey point. This makes the use of this method particularly difficult in offshore surveying. An offshore method for a magnetotelluric survey is described in U.S. Pat. No. 4,210,869, but this method relies upon acoustic methods of measuring the position and orientation of the electrodes on the sea floor. The accuracy of these measurements is limited, making the interpretation of the survey results unreliable.
A further disadvantage of conventional magnetotelluric methods is that measurements of the magnetic field at several points in the survey area is required, unless the area to be surveyed is small. In addition, the vertical component must be ascertained at each E measurement point in order to distinguish between the apparent E parallel and E perpendicular components. As explained above, if this distinction is not made, there is a serious likelihood that errors will be made in interpreting the data obtained.